IpAddressGenerationFailure No more IP addresses
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
neutron |
Invalid
|
Undecided
|
Unassigned |
Bug Description
Hi Guys,
When a user create a wrong network x.x.x.x/32 or subnet is full, neutron generate this error:
2014-12-22 06:31:48.970 6318 ERROR neutron.
2014-12-22 06:31:48.970 6318 TRACE neutron.
When this happens neutron restart all dnsmasq services and all VMs fails to renew DHCP lease:
Dec 22 16:37:25 mariadb2 dhclient: DHCPREQUEST of 10.29.81.115 on eth0 to 10.29.81.3 port 67 (xid=0x1997b37a)
Dec 22 16:37:25 mariadb2 dhclient: DHCPNAK from 10.29.81.3 (xid=0x1997b37a)
Dec 22 16:37:25 mariadb2 dhclient: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 3 (xid=0x62a70cd4)
Dec 22 16:37:25 mariadb2 dhclient: DHCPREQUEST of 10.29.81.115 on eth0 to 255.255.255.255 port 67 (xid=0x62a70cd4)
Dec 22 16:37:25 mariadb2 dhclient: DHCPOFFER of 10.29.81.115 from 10.29.81.3
Dec 22 16:37:25 mariadb2 dhclient: DHCPACK of 10.29.81.115 from 10.29.81.3
Dec 22 16:37:25 mariadb2 dhclient: bound to 10.29.81.115 -- renewal in 50 seconds.
In an environment with many traffic this problem create a packet loss and the dnsmasq service are overworked
Do you have any idea to resolve that?
This looks like support request, not a bug report. Please ask openstack ML, ask.openstack.org, or corresponding irc channel.